Background/Question/Methods   Roads directly contribute to the decline of animal populations due to road killings, and landscape fragmentation, eventually leading to patches that are too small to support some populations. Fences provide a cheap measure to mitigate road killings, although they may also adversely affect a population by isolating patches and stopping gene flow between populations. Here, we develop an analytical and simulation approach to analyze the impact of fences on the survival of a given population. Our method is based on Skellam's model of diffusion, and we use difference equations and simulations to deal with problems that are not analytically tractable. Our approach is to, first, determine the minimum unfenced area for a population characterized by given growth and diffusion rates to survive, and, second, to study what the characteristics of the fencing should be, to allow for the survival of a population within smaller, fenced patches.

Results/Conclusions   Our simulations show that the total perimeter and the spatial layout of the fences play a major role in the survival or demise of populations. In particular, we show that the spatial arrangement of fences strongly affects both the equilibrium density and spatial distribution of species. This factor can even lead to the extinction of a population in some patches and survival in others. Our methods provide a tool to assess the effectiveness of fences as a mitigation measure to prevent the decline and extinction of populations in an anthropogenically disturbed landscape.